Date post: | 28-Nov-2023 |
Category: |
Documents |
Upload: | independent |
View: | 0 times |
Download: | 0 times |
Chemistry & Biology
Article
Amide N-Glycosylation by Asm25,an N-Glycosyltransferase of AnsamitocinsPeiji Zhao,1,5 Linquan Bai,2,5,* Juan Ma,1 Ying Zeng,1 Lei Li,2 Yirong Zhang,2 Chunhua Lu,3 Huanqin Dai,1
Zhaoxian Wu,3 Yaoyao Li,3 Xuan Wu,3 Gang Chen,1 Xiaojiang Hao,1 Yuemao Shen,1,3,* Zixin Deng,2
and Heinz G. Floss4
1State Key Laboratory of Phytochemistry and Plant Resources in West China, Kunming Institute of Botany, Chinese Academy of Sciences,Kunming, Yunnan 650204, China2Laboratory of Microbial Metabolism, College of Life Science and Biotechnology, Shanghai Jiaotong University, Shanghai 200240, China3Key Laboratory of the Ministry of Education for Cell Biology and Tumor Cell Engineering, School of Life Sciences, Xiamen University, Xiamen,
Fujian 361005, China4Department of Chemistry, Box 351700, University of Washington, Seattle, WA 98195-1700, USA5These authors contributed equally to this work.
*Correspondence: [email protected] (L.B.), [email protected] (Y.S.)
DOI 10.1016/j.chembiol.2008.06.007
SUMMARY
Ansamitocins are potent antitumor maytansin-oids produced by Actinosynnema pretiosum. Theirbiosynthesis involves the initial assembly of a macro-lactam polyketide, followed by a series of postpoly-ketide synthase (PKS) modifications. Three ansami-tocin glycosides were isolated from A. pretiosumand fully characterized structurally as novel ansami-tocin derivatives, carrying a b-D-glucosyl groupattached to the macrolactam amide nitrogen in placeof the N-methyl group. By gene inactivation andcomplementation, asm25 was identified as the N-gly-cosyltransferase gene responsible for the macrolac-tam amide N-glycosylation of ansamitocins. Soluble,enzymatically active Asm25 protein was obtainedfrom asm25-expressing E. coli by solubilizationfrom inclusion bodies. Its optimal reaction condi-tions, including temperature, pH, metal ion require-ment, and Km/Kcat, were determined. Asm25 alsoshowed broad substrate specificity toward otheransamycins and synthetic indolin-2-ones. To thebest of our knowledge, this represents the firstin vitro characterization of a purified antibiotic N-gly-cosyltransferase.
INTRODUCTION
Ansamitocins are maytansinoids of microbial origin, produced
by Actinosynnema pretiosum ssp. auranticum ATCC 31565.
They differ structurally from their higher-plant congeners, such
as the parent compound maytansine, only by carrying simpler
acyl groups at C-3 (Figure 1A; Higashide et al., 1977). Like the
plant maytansinoids, ansamitocins show extremely potent
cytotoxicity against various tumor cells (Ootsu et al., 1980).
Structurally, the maytansinoids belong to the ansamycin family
of polyketide macrolactams (Rinehart and Shield, 1976), exem-
plified by the antitubercular agent rifamycin and the antitumor
Chemistry & Biology 15, 8
agent geldanamycin. Their molecular architecture is character-
ized by an aromatic chromophore with an aliphatic chain (ansa
chain) connected back to a nonadjacent position through an
amide linkage. Despite carrying either benzenic (ansamitocins,
geldanamycin) or naphthalenic (rifamycin, naphthomycin) chro-
mophores, genetic and feeding experiments have revealed that
all the ansamycins share the same polyketide starter unit,
3-amino-5-hydroxybenzoic acid (AHBA) (Ghisalba and Nuesch,
1981; Hatano et al., 1982; Kibby et al., 1980). The ansamitocin
biosynthetic genes were cloned from a cosmid library of A. pre-
tiosum genomic DNA using the AHBA synthase gene rifK (Kim
et al., 1998) from the rifamycin biosynthetic gene cluster as
probe (Yu et al., 2002). Notably, two AHBA synthase genes
were found, and sequencing of the surrounding DNA localized
the genes expected to be required for ansamitocin biosynthesis
in two clusters separated by 30 kb of nonessential DNA
(Figure 1B). Bioinformatic analysis, gene inactivation, and ex-
pression experiments demonstrated that cluster I contains
most of the ansamitocin biosynthetic genes, whereas cluster II
carries only four of the seven genes required for AHBA formation
(Figure 1B; Yu et al., 2002). AsmA-D are four large open-reading
frames (ORFs) encoding the loading domain and seven chain-
elongation modules of a type I multifunctional polyketide syn-
thase (PKS). Gene asm9, located immediately downstream of
asmD, encodes the amide synthase responsible for macrolac-
tamization of the nascent polyketide chain and release of proan-
samitocin from the PKS (Kato et al., 2002; Yu et al., 2002).
As in other microbial secondary metabolites, post-PKS modi-
fications are important for the biological activity of the ansamito-
cins (Cassady et al., 2004). Six modifying reactions are required
to turn proansamitocin into ansamitocins, and cluster I contains
candidate genes for each of these. Their functions and the order
in which they act were determined by in vivo gene inactivation,
feeding experiments with accumulated compounds, and
in vitro enzymatic analysis of heterologously expressed gene
products. This defined a predominant pathway from proansami-
tocin to ansamitocins, consisting of chlorination (Asm12), carba-
moylation (Asm21), O-methylation (Asm7), acylation (Asm19),
epoxidation (Asm11), and N-methylation (Asm10) (Moss et al.,
2002; Spiteller et al., 2003). Notably, the only ansamitocin prod-
ucts of the asm10 deletion mutant are macrolactam amide
63–874, August 25, 2008 ª2008 Elsevier Ltd All rights reserved 863
Chemistry & Biology
Amide N-Glycosylation
Figure 1. The Post-PKS Modifications of N-Desmethylansamitocins and the Ansamitocin Biosynthetic Gene Cluster
(A) Chemical structures of ansamitocins (APs), N-desmethylansamitocins (PNDs), and ansamitocinosides (AGPs). The putative scheme for the conversion of
PNDs to APs and AGPs by Asm10 and Asm25, respectively.
(B) The ansamitocin biosynthetic gene cluster. The N-glycosyltransferase gene, asm25, and the N-methyltransferase gene, asm10, are indicated by a circle in
Cluster I. AHBA: 3-amino-5-hydroxybenzoic acid; MM-ACP: methoxy-malonyl-ACP.
N-desmethylansamitocins (PNDs) (Figure 1A), indicating that
N-methylation is the terminal step in the post-PKS modifications
of ansamitocin biosynthesis (Spiteller et al., 2003).
Structural variations of the ansamitocins reside mainly in the
C-3 acyl moiety. Other minor components with hydroxyl func-
tions at C-15 or the C-14 methyl group, or lacking the N-methyl
group, were also found in A. pretiosum (Izawa et al., 1981).
Recently, from the cultures of A. pretiosum on solid ISP2 me-
dium, we have isolated a series of novel N-glycosylated ansami-
tocins with a b-D-glucosyl residue attached to the macrolactam
amide nitrogen instead of the N-methyl group (Figure 1A). Ansa-
mitocinosides P-1 and P-2 (AGP-1, AGP-2), the main compo-
nents, were fully characterized structurally (Lu et al., 2004; Ma
et al., 2007). In the present work, we show through gene inacti-
vation and complementation the gene encoding the unique am-
ide N-glycosyltransferase (N-Gtf) to be asm25, and we report the
overexpression, purification, and characterization of Asm25.
Soluble active Asm25 protein was purified from inclusion bodies
formed in asm25-expressing E. coli, which is unprecedented in
studies of antibiotic N-Gtfs. Additionally, the substrate range of
864 Chemistry & Biology 15, 863–874, August 25, 2008 ª2008 Elsev
Asm25 was probed with other ansamycins and synthetic lac-
tams (indolin-2-one derivatives).
RESULTS
Isolation and Structure Elucidationof Ansamitocinoside P-3Previously, we reported the isolation of two unique ansamitocin
N-glycosides, macrolactam amide N-desmethyl-N-b-D-gluco-
pyranosylansamitocins P-1 and P-2, named as ansamitocino-
sides P-1 (AGP-1) and P-2 (AGP-2) from A. pretiosum ATCC
31565 (Lu et al., 2004; Ma et al., 2007), and detected additional
ansamitocinosides by LC-MS. In the present work, one more
glycoside of a macrolactam amide N-desmethylansamitocin,
namely ansamitocinoside P-3 (AGP-3), was isolated from the
ethyl acetate extract of A. pretiosum cultivated on solid ISP2 me-
dium by improved preparative TLC methods (Ma et al., 2007).
The molecular formula of AGP-3 was determined to be
C37H51N2O14ClNa (m/z 805.2917 [M + Na]+, calculated:
805.2926) based on HR-ESI-MS data. The 13C-NMR and DEPT
ier Ltd All rights reserved
Chemistry & Biology
Amide N-Glycosylation
Figure 2. Inactivation of asm25 and Complementation of the Mutant
(A) Schematic representation of the 465 bp deletion within asm25 from the A. pretiosum genome. In shuttle plasmid pJTU478, a truncated asm25 was generated
by linking the 2.7 kb and 3.4 kb genomic fragments originally flanking the deleted 465 bp region. S, Sacl; P, Pstl.
(B) PCR analysis of wild-type A. pretiosum and mutant BLQ11. Whereas wild-type A. pretiosum gave a 1.60 kb PCR-amplified product, mutant BLQ11 yielded
a 1.10 kb product as expected.
(C) ESI-MS profiles of the ansamitocin products of the wild-type, asm25 deletion mutant BLQ11, and asm25-complemented strain BLQ15 cultured on solid ISP2
medium.
spectra of AGP-3 show 37 carbon signals, including 7 methyls, 4
methylenes, 16 methines, and 10 quaternary carbons (see Table
S1A available online). But in the 1H-NMR spectra, the proton sig-
nal attributed to CH3N at d 3.18�3.22 in ansamitocin P-3 (AP-3)
(Kupchan et al., 1977) was missing in AGP-3, indicating the ab-
sence of the N-methyl group in AGP-3. In addition, one more
six-carbon unit (d 85.3, 71.6, 80.0, 71.5, 79.5, 63.3) was assigned
to be a hexosyl moiety. The sugar was determined to be a b-D-
glucopyranosyl moiety based on the unambiguous NMR assign-
ments. Particularly, the JH-1’’,-2’’ = 9.4 Hz indicated a b-glucosidic
linkage. The HMBC correlations between the anomeric proton
at d 5.55 (H-100) and the carbons at d 172.6 (C-1) and d 138.3 (C-18)
revealed the linkage of the glucose unit to the aglycone, N-des-
methylansamitocin P-3 (PND-3), via the macrolactam amide
nitrogen. Therefore, AGP-3 was determined to be N-des-
methyl-N-b-D-glucopyranosylansamitocin P-3 (Figure 1A).
asm25 Is Involved in the Formationof AnsamitocinosidesEarly work had established methylation of the macrolactam am-
ide nitrogen as the terminal step in post-PKS modifications and
identified PND-3 as the last intermediate in ansamitocin biosyn-
thesis (Spiteller et al., 2003). The glycosylation of the same amide
Chemistry & Biology 15, 8
nitrogen could be an alternative step to the methylation catalyzed
by the N-methyltransferase Asm10. Sequence analysis of the
asm gene cluster had identified one gene of unknown function,
asm25, with high sequence homology to glycosyltransferase
(Gtf) genes (Walsh et al., 2003), which at the time had been spec-
ulated to be part of a glycosylation/deglycosylation excretion
system (Figure 1B; Yu et al., 2002). To investigate whether
asm25 is responsible for the N-glycosylation of PNDs, this gene
was inactivated through a 465 bp in-frame deletion from 429 bp
to 894 bp (Figure 2A). The resulting mutant strain BLQ11 was
confirmed by PCR amplification using the total genomic DNA
as template. On the electrophoretic gel, BLQ11 gave the 1.1 kb
expected product, whereas the wild-type showed a 1.6 kb prod-
uct (Figure 2B). The extracts of both BLQ11 and wild-type strains
grown on solid ISP2 medium were analyzed by LC-ESI-MS,
monitoring at the quasimolecular ion peaks of ansamitocins
(APs), ansamitocinosides (AGPs), and N-desmethylansamitocins
(PNDs), respectively. The production of AGP-1, AGP-2, AGP-3,
and AGP-4 (at m/z 755.3 [M + H]+, 769.3 [M + H]+, 783.3 [M + H]+,
and 797.3 [M + H]+, respectively) was observed in the wild-type
but abolished in BLQ11 (Figure 2C).
Further support for the involvement of asm25 in the N-glyco-
sylation at the macrolactam amide nitrogen came from
63–874, August 25, 2008 ª2008 Elsevier Ltd All rights reserved 865
Chemistry & Biology
Amide N-Glycosylation
Figure 3. Properties of the Renatured
Asm25
(A) A representative Coomassie brilliant blue-
stained SDS-PAGE gel loaded with molecular
weight markers (lane 1), renatured Asm25 (lane
2), inclusion bodies (lane 3), soluble protein (lane
4), and cell-free extract (lane 5) fractions.
(B) Temperature optimum for the macrolactam
amide N-glycosylation of PND-3 catalyzed by
Asm25.
(C) Determination of pH optimum for the macrolac-
tam amide N-glycosylation of PND-3 catalyzed by
Asm25. Asm25 was assayed in 50 mM Na-ace-
tate/citric acid buffer ranging from pH 4.0 to 7.0
(open column), and 50 mM Tris-HCl buffer ranging
from pH 7.0 to 9.0 (solid column).
(D) Effect of metal ions on the activity of Asm25.
Asm25 was assayed in the absence (0 mM, green)
or presence of various metal ions at 0.1 (blue), 1.0
(pink), and 10.0 mM (black) concentrations except
for the concentration of Mg2+ used at 10 (green
and blue), 50 (pink), and 100 mM (black). The pro-
ductions of AGP-3 were quantitated by LC-MS
with authentic AGP-3 as external standard (see
Supplemental Data).
complementation of mutant BLQ11. An intact copy of asm25
under the control of the PermE* promoter was introduced
into BLQ11 by conjugation using an integrative vector
pJTU813, which generated a derivative named as BLQ15.
LC-MS analysis of the extract of BLQ15 showed a chromato-
gram nearly identical to that of the wild-type; that is, the pro-
duction of AGPs was fully restored by expression of the cloned
asm25 (Figure 2C).
In Vitro N-Glycosylation of PNDs by Asm25To verify that Asm25 indeed catalyzes the glycosylation at the
macrolactam amide nitrogen of ansamitocins, the asm25 gene
was amplified by PCR from the A. pretiosum genome and cloned
into plasmid pRSET-B for heterologous expression. The expres-
sion plasmid pJTU812 was transformed into E. coli BL21(DE3)-
plysE and induced at different temperatures and/or with different
concentrations of isopropyl-b-D-thiogalactopyranoside (IPTG).
Asm25 was successfully overexpressed as an N-terminally
His6-tagged fusion protein as detected by SDS-PAGE. The esti-
mated molecular weight of Asm25 recombinant protein is 45 kDa
from the SDS-PAGE (Figure 3A), and its theoretical molecular
weight is 45679.85 Da.
The yield of recombinant protein was highest when the cells
were grown at 37�C at different concentrations of IPTG, but as
shown in Figure 3A, almost all of it was present as inclusion
bodies. In contrast, cells growing at 16�C or 20�C induced
with 0.8 mM IPTG produced only minute amounts of soluble
Asm25, which showed weak catalytic activity when incubated
with PND-3 and UDP-Glc (data not shown). After many unsuc-
cessful attempts to purify the soluble recombinant Asm25 by
metal ion affinity chromatography, the inclusion bodies were
chosen for the preparation of Asm25. Soluble His6-Asm25 pro-
tein with the best enzymatic activity was obtained from the
inclusion bodies by washing with 2.0 M urea and 0.5% Triton,
solubilization in 8.0 M urea, and finally dialysis against renatur-
ation buffer IV with glycerol and glycine (see Experimental Pro-
866 Chemistry & Biology 15, 863–874, August 25, 2008 ª2008 Elsev
cedures). The solution of Asm25 (1.6 mM) in buffer IV with 30%
glycerol and 1% bovine serum albumin (BSA) was used for the
enzyme assays.
Based on the structures of the ansamitocinosides and the
known pathway of ansamitocin biosynthesis (Spiteller et al.,
2003), it was assumed that PNDs are the aglycone substrates
for Asm25. Using as substrate PND-3 isolated from the asm10
mutant grown on solid ISP2 medium, the enzymatic properties
of Asm25 were examined. The reactions were performed with
10 ml PND-3 (20 mM in DMSO), 10 ml UDP-Glc (40 mM; Sigma,
St. Louis, MO), and 0.1 ml basal reaction solution (1.6 mM
Asm25, 50 mM Tris-HCl [pH 8.0], 30% glycerol, 1% BSA,
0.1 M NaCl, 0.1 M glycine, 0.5 mM EDTA, 1.0 mM DDT, and
10 mM MgCl2). The reaction product was identified as AGP-3
by MS (Figure S2B), NMR, and HPLC comparison (Figures
S2H–M and Table S1B), and quantitated by LC-MS with au-
thentic AGP-3. Based on the reaction temperature profile
from 0�C to 50�C, Asm25 purified from the inclusion bodies
was optimally active at 37�C and exhibited more than 70%
of its maximal activity between 35�C and 42�C (Figure 3B).
The pH profile was determined at 37�C in 50 mM Na-ace-
tate/citric acid buffer (pH 4.0–7.0) and 50 mM Tris-HCl buffer
(pH 7.0–9.0). Asm25 showed optimal activity at pH 7.5 and
more than 70% of maximal activity in the pH range of 6.0–
8.0 (Figure 3C).
Because many enzymes that bind dinucleotide substrates re-
quire divalent, oxophilic cations, such as Mg2+ or Mn2+, we
examined the effects of metal ions on the activity of Asm25.
The reactions were performed in the basal reaction solution con-
taining 1.54 mM PND-3 and 3.08 mM UDP-Glc for 4 h at 37�C.
Enzymatic activity improved with the addition of Li+, Zn2+, or
Mn2+ at a concentration of 0.1 mM. Negligible effects on the cat-
alytic activity were observed as the concentrations of Li+ and
Mn2+ increased to 10 mM. However, the activity was partially in-
hibited by 0.1 mM Hg2+ or Cu2+, and almost completely inhibited
by 1 mM Hg2+, Zn2+, or Cu2+ (Figure 3D).
ier Ltd All rights reserved
Chemistry & Biology
Amide N-Glycosylation
Kinetics of Renatured Asm25 with UDP-Glcand PND-3 as CosubstratesKinetic studies of the reaction catalyzed by the refolded
Asm25 were performed using UDP-Glc as sugar donor and
PND-3 as sugar acceptor. When the initial rate was measured
by varying the concentrations of PND-3 at different fixed con-
centrations of UDP-Glc, a set of intersecting lines were ob-
tained by regression analysis of the data at each UDP-Glc
concentration (Figure 4A). Double reciprocal plots of the initial
rate data are shown in Figure 4A. The lines in the double recip-
rocal plots converged, though not to a single point, implying
that Asm25 obeys a ternary complex mechanism in which
both substrates bind prior to product release. In general, two
basic types of mechanisms, ternary complex and ‘‘ping-
pong’’ mechanism, can be deduced from kinetic data based
on the lines in double reciprocal plots. Lines converging to
a single point indicate a ternary complex mechanism, whereas
parallel lines suggest a ping-pong mechanism. The lines in Fig-
ure 4A do not converge to a single point; however, they show
the tendency to converge, which differs from the parallel pat-
tern. Therefore, a ternary complex mechanism can be con-
cluded for Asm25. This was also seen in the Arabidopsis in-
dole-3-acetic acid Gtf, which has a mechanism similar to
Asm25 (Jackson et al., 2001). The slope replot does not
pass through the origin, indicating that the glycosylation is
not of a rapid equilibrium-ordered mechanism (Figure 4B).
The reaction thus proceeds through a ternary enzyme-sub-
strate complex (Lairson et al., 2008; Marangoni, 2003; Price
and Stevens, 1982). The recombinant Asm25 exhibited
Michaelis-Menten kinetics. The Vmax and Km values were de-
termined by classical initial rate experiments and calculated
from Lineweaver-Burk plots (Price and Stevens, 1982). The
Km for PND-3 and UDP-Glc were 17.8 mM and 87.3 mM,
respectively, with kcat of 1.3 $ 10�2 $ s�1.
Figure 4. Kinetic Characterization of Asm25
(A) Initial reaction rates were determined with
PND-3 at 20, 30, 40, and 50 mM and UDP-Glc at
0.2, 0.5, 0.75, 1.0, and 2.0 mM, respectively.
Mean values of three independent experiments
with SD indicated by error bars.
(B) Secondary plot of intercept versus 1/[PND-3]
and slope versus 1/[PND-3].
Substrate Specificity of Asm25The aglycone specificity of Asm25
was evaluated using six maytansinoids
(see the LC-MS profiles in Figure S1),
including PND-1 to -4, N-desmethyl-
desepoxymaytansinol (DDM) and N-des-
methyl-desepoxyansamitocin P-1 (DDP-
1) to react with UDP-Glc, respectively.
The glycosylated products, AGP-1, -2,
-3, -4, corresponding to the four PNDs,
were detected by LC-MS (Figure 5A,
panels 1–4) and HPLC (Figure S3). By
comparison of the relative signal intensi-
ties of the product ions, LC-MS analysis
indicated that the glycosylation of PND-2
to AGP-2 gave the highest yield. This is consistent with the
competition reaction using equimolar amounts of all four
PNDs, which showed that among these PND-2 was the preferred
substrate of the enzyme (Figure 5A, panel 5). In contrast, no con-
version of DDM to its corresponding glycoside was detected
(data not shown). A glycosylation product of DDP-1 was de-
tected, but the conversion efficiency was extremely low (data
not shown). Therefore, among the substrates tested, PND-2
was the best sugar acceptor for the glycosylation reaction cata-
lyzed by Asm25.
The sugar specificity of Asm25 was probed using PND-3 and
four UDP-sugars, UDP-Glc, UDP-galactose, UDP-N-acetyl-glu-
cosamine and UDP-glucuronic acid, and ADP- and GDP-Glc as
substrates. A glycosylated product could only be detected by
LC-MS in the assay with UDP-Glc as sugar donor (Figure 5B
and Figure S4). Because no glycosylated products were formed
in the other five reactions, UDP-galactose, UDP-N-acetyl-
glucosamine, UDP-glucuronic acid, ADP- or GDP-Glc cannot
replace UDP-Glc in the Asm25-catalyzed glycosylation of
PND-3. Therefore, we conclude that Asm25 catalyzes the
in vitro glycosylation of PNDs at the macrolactam amide nitrogen
using UPD-Glc as the sole sugar donor.
In Vitro Glycosylation of Other Ansamycinsand Synthetic Lactams by Asm25To evaluate the substrate range of Asm25 further, five ansamy-
cins—geldanamycin-derived lebstatin and 17-O-desmethylleb-
statin (17-DMLB) (Figure 6; Stead et al., 2000), rifamycin A (Sensi
et al., 1959), rifampicin (Morisaki et al., 1995), naphthomycin A
(Lu and Shen, 2007)—and two synthetic lactams (indolin-2-one
derivates HC-14 and HC-52; see Supplemental Data) were
used as sugar acceptors at 8.33 mM in standard assays with
UDP-Glc. Rifamycin A, rifampicin, and naphthomycin A were
not converted to the corresponding glycosides (data not shown).
Chemistry & Biology 15, 863–874, August 25, 2008 ª2008 Elsevier Ltd All rights reserved 867
Chemistry & Biology
Amide N-Glycosylation
Figure 5. Substrate Specificity of Asm25 for Different Ansamitocin-Derived Aglycones and Sugar Donors
(A) ESI-MS spectra of the reaction products of Asm25 with different ansamitocin-derived aglycones. Panels 1–4: Incubations with different PNDs. Panel 5:
Incubation with equimolar mixture of four PNDs. All reactions were carried out in standard assays, at least in duplicate (see Supplemental Data).
(B) ESI-MS spectra of the reaction products of Asm25 with different UDP-sugar donors. For assaying different sugar donors, 10 ml of sugar nucleotide (0.15 mM
final concentration) and 10 ml PND-3 (0.04 mM final concentration) were added to 0.1 ml basal reaction solution for a 0.12 ml final volume. The reactions were
terminated after 2 h at 37�C and the products were detected by LC-MS. Each assay was carried out at least in duplicate.
However, lebstatin, 17-DMLB, HC-14, and HC-52 were glycosy-
lated by Asm25. In the LC-MS spectra, quasimolecular ion peaks
were seen for lebstatin at m/z 571.4 [M + Na]+ and for the ex-
868 Chemistry & Biology 15, 863–874, August 25, 2008 ª2008 Elsev
pected glycosylated product at m/z 733.4 [M + Na]+. Similarly,
in the LC-MS spectra of the incubation with 17-DMLB, the glyco-
sylated product was detected at m/z 719.4 [M + Na]+. Moreover,
ier Ltd All rights reserved
Chemistry & Biology
Amide N-Glycosylation
Figure 6. ESI-MS Spectra of the Glycosylation on Selected Nonansamitocin Aglycones by Asm25
17-DMLB, lebstatin, HC-14, and HC-52 were used as sugar acceptors at 8.33 mM in standard assays with UDP-Glc, respectively (see Supplemental Data).
Chemistry & Biology 15, 863–874, August 25, 2008 ª2008 Elsevier Ltd All rights reserved 869
Chemistry & Biology
Amide N-Glycosylation
the corresponding glycosylated products of HC-14 and HC-52
were observed at m/z 366.5 [M + H]+ and 582.9 [M + Na]+,
respectively (Figure 6).
Antifungal and Antitumor Activitiesof PND-3, AP-3, and AGP-3AP-3, AGP-3, and PND-3 were tested for antifungal activity
against yeast Filobasidium uniguttulatum IFO 0699. AP-3 and
PND-3 showed remarkable activity against F. uniguttulatum at
50–200 ng/disc, but AGP-3 showed similar activity only at higher
concentration (8–10 mg/disc; Figure 7A). This assay indicated
that AP-3 possesses the strongest antifungal activity, followed
by PND-3 and AGP-3.
The antitumor activities of AP-3, AGP-3, and PND-3 against
the HepG2 cell line were assayed by the MTT method (Mos-
mann, 1983). AP-3 and PND-3 showed similar activities with
IC50 of 0.2 mg/ml and 0.1 mg/ml, respectively, which is consistent
with values reported for maytansinoids (Sneden and Beemster-
boer, 1980). AGP-3 was 10 times less active than AP-3
(Figure 7B). However, the in vivo antitumor activity of AGP-3
would still be worth testing because some maytansinoids have
been reported to have low activity during in vitro test systems
but a dramatic in vivo antitumor activity (Cassady et al., 2004).
DISCUSSION
The post-PKS modification reactions in the biosynthesis of ansa-
mitocins are diverse, including chlorination, carbamoylation,
O-methylation, acylation, epoxidation, and N-methylation. Nev-
ertheless, the structural variations among the naturally occurring
Figure 7. Antifungal and Antitumor Activi-
ties of AP-3, PND-3, and AGP-3
(A) Antifungal activity against F. uniguttulatum IFO
0699 of AP-3, PND-3, and AGP-3.
(B) The antitumor activities of AP-3, PND-3, and
AGP-3 against HepG2 cells. The IC50 of AP-3,
PND-3, and AGP-3 were calculated to be 0.2,
0.1, and 2.0 mg/ml, respectively.
ansamitocins are limited mainly to the
C-3 acyl moieties and to the omission of
one or two of the post-PKS modification
steps. Our isolation from A. pretiosum of
three novel ansamitocin amide N-glyco-
sides, AGP-1, AGP-2, and AGP-3
(Figure 1A), revealed a novel post-PKS
modification reaction not heretofore
encountered in ansamitocin biosynthesis
(Lu et al., 2004; Ma et al., 2007).
Bioinformatic analysis of the asm gene
cluster had revealed the presence of
asm25 encoding a putative Gtf, but its
function was obscure (Yu et al., 2002).
Asm25 showed highest sequence simi-
larity to CalG1 in the calicheamicin
biosynthetic gene cluster cloned from
Micromonospora echinospora ssp. cali-
chensis (27% identity, 42% similarity)
(Ahlert et al., 2002). In the present work, the Asm25 protein has
now been characterized as a dedicated tailoring Gtf for ansami-
tocins. Gtfs (EC 2.4.x.y) constitute a large family of enzymes that
are involved in the biosynthesis of oligosaccharides, polysac-
charides, glycoconjugates, and other natural products. Glyco-
syltransfers are nucleophilic replacement reactions between
C-1 of nucleotide-activated sugars and aglycones that usually
carry nucleophilic hydroxyl substituents. However, less fre-
quently, amines and nucleophilic carbons can also serve as
sugar acceptors, such as in plant xenobiotic metabolism by
the bifunctional O- and N-Gtf UGT72B1 (Brazier-Hicks et al.,
2007), and in the biosynthesis of antibiotics—e.g., rebeccamycin
(Sanchez et al., 2002), urdamycin (Hoffmeister et al., 2000), and
enterobactin (Fischbach et al., 2005). Many N-glycosides are
found among the indolocarbazole antibiotics (Sanchez et al.,
2006). The first N-Gtf-encoding gene (ngt/rebG) was cloned
from Lechevalieria aerocolonigenes ATCC 39243 (Hyun et al.,
2003; Nishizawa et al., 2005; Ohuchi et al., 2000; Onaka et al.,
2003a, 2003b; Sanchez et al., 2002). It is involved in the biosyn-
thesis of the indolocarbazole alkaloid rebeccamycin. Later, three
more N-Gtf genes, staG (Onaka et al., 2002; Salas et al., 2005),
atmG (Gao et al., 2006), and inkG (Kim et al., 2007), were
identified.
So far, the biochemical characteristics and/or substrate spec-
ificity profiles of the four known N-Gtfs involved in N-glycosyla-
tions of indolocarbazole antibiotics have been investigated
only by in vivo experiments due to the difficulty of obtaining sol-
uble active proteins (Sanchez et al., 2005; Zhang et al., 2006).
The expression of the rebG gene in E. coli and in Streptomyces
lividans led to RebG overproduction as inclusion bodies. All
870 Chemistry & Biology 15, 863–874, August 25, 2008 ª2008 Elsevier Ltd All rights reserved
Chemistry & Biology
Amide N-Glycosylation
efforts to obtain soluble and functional RebG were unsuccessful.
As the best outcome, RebG was partially solublized by fusion
with maltose binding protein or by coexpression with DnaK/
DnaJ chaperones; however, this soluble RebG was inactive dur-
ing in vitro assays (Zhang et al., 2006). Just as encountered in the
heterologous expression of RebG, StaG, and AtmG, the overpro-
duction of Asm25 in E. coli was successful, but produced inclu-
sion bodies. Because in the previous work many strategies had
been attempted without any success (Zhang et al., 2006), we
decided to focus on refolding Asm25 after purification from the
inclusion bodies under denaturing conditions, as this seemed
the most straightforward approach.
Our extensive efforts in refolding Asm25 from denaturated
protein eventually produced soluble active enzyme in reasonable
yield (�4.38 mg/l of induced cells). When inclusion bodies were
purified, 2 M urea was added in renaturation buffer IIa to further
purify Asm25 before solubilizing, and which also removed non-
specifically adsorbed proteins. To refold Asm25, we used a pro-
cedure similar to that described previously (Yang et al., 2004),
except that glycerol, which has long been known to increase pro-
tein stability (Gekko and Timasheff, 1981), and glycine were
added to renaturation buffer IV. After adding 0.1 M glycine, the
activity of Asm25 was increased dramatically (data not shown).
The use of low molecular-weight additives such as L-arginine,
sugars, glycerol, and sarcosine during the refolding process
often helps in improving the yield of active proteins from inclu-
sion bodies (Clark, 1998). These additives influence both the sol-
ubility and stability of the unfolded protein, folding intermediates,
and the fully folded protein. With the soluble Asm25 available, we
characterized the N-glycosylation reaction in vitro.
As anticipated, we found that UDP-Glc is the sugar donor and
PNDs are the sugar acceptors for the Asm25-catalyzed glycosyl-
ation to generate AGPs (Figure 1A). Using UDP-Glc and PND-3
as cosubstrates, we investigated the kinetics of renatured
Asm25 and found that the enzyme catalyzes the N-glycosylation
of PND-3 via the formation of a ternary complex (Marangoni,
2003; Price and Stevens, 1982). Ternary complex formation is
entirely in agreement with the theory—proposed on the basis
of structural and sequence homology studies of Gtfs—that
UDP-Glc transferases possess two major functional domains
(Mackenzie, 1990). Each of these domains is thought to contain
a binding site for one of the cosubstrates, enabling both sub-
strate molecules to bind to the enzyme simultaneously. This ter-
nary complex sequential mechanism is in agreement with the
limited studies carried out on other Gtfs (Jackson et al., 2001;
Mulichak et al., 2001).
The evaluation of the substrate specificity of Asm25 for various
sugar acceptors showed that all four PNDs tested (Figure 1A) are
used as substrates, but that the structure of the C-3 acyl group
modulates the efficiency of conversion. Both in individual incu-
bations and in a competition experiment with all four com-
pounds, PND-2 was found to be the preferred substrate. This
observation suggests that the C-3 acyl group may serve as
a binding site in a hydrophobic interaction between PNDs and
the aglycone binding pocket of Asm25. Additionally, this result
is consistent with the fact that AGP-2 was the most abundant
product in the fermentation of A. pretiosum on solid ISP2
medium (data not shown). Two N-desmethylansamitocins lack-
ing the 4,5-epoxy function, DDM and DDP-1 (Figure S1), were
Chemistry & Biology 15, 8
extremely poor substrates, if at all. This further supports the
speculation that the macrolactam amide N-methylation and/or
N-glycosylation is the final step in the post-PKS modifications
of ansamitocins (Spiteller et al., 2003). However, it is rather sur-
prising given the finding that two other benzenic ansamycins, the
geldanamycin derivatives lebstatin and 17-DMLB (Figure 5A), do
serve as substrates. The site of glycosylation has not been deter-
mined with these substrates, but is assumed to be the macrolac-
tam amide nitrogen. Three naphthalenic ansamycins, rifamycin
A, rifampicin, and naphthomycin A, however, were not glycosy-
lated by Asm25, although two synthetic lactams, HC-14 and
HC-52 (Figure 6), were converted to glycosides with modest
efficiency. Though these results do not provide a clear picture
of the aglycone binding site, they do indicate considerable pro-
miscuity of the enzyme toward sugar acceptor substrates.
The relative promiscuity of Asm25 with respect to the aglycone
substrate contrasts with the rather high specificity of the enzyme
for just UDP-Glc as the sugar donor. The UDP derivatives of
galactose, glucuronic acid, and N-acetylglucosamine, and
ADP- and GDP-Glc, were not used as substrates. Similarly, it
has been reported that no sugar moieties other than D-glucose
are utilized for the glycosylations catalyzed by RebG (Onaka
et al., 2003a, 2003b; Sanchez et al., 2002, 2005; Zhang et al.,
2006). Conversely, StaG accepts a variety of sugar moieties, in-
cluding L-rhamnose, L-olivose, L-digitoxose, and D-olivose dur-
ing in vivo biotransformations (Salas et al., 2005). An examination
of the structures in the PDB database shows that the conforma-
tions of UDP-hexose substrates bound to enzymes can vary sig-
nificantly. The variations occur largely around the diphosphate
linkage, which has shallow rotational barriers and the ability to
adopt a large number of isoenergetic conformations. Thus, the
nucleoside and hexose portions of the substrate can be pre-
sented in different orientations depending on the enzyme,
reflecting large differences in the substrate binding pockets.
SIGNIFICANCE
Ansamitocins are the microbial versions of natural maytan-
sinoids, a family of 19-membered macrocyclic lactams
with extraordinary cytotoxic and antineoplastic activities,
which are currently undergoing clinical evaluation as anti-
body-conjugated antitumor drugs (Widdison et al., 2006).
Their biosynthesis, catalyzed by the products of the asm bio-
synthetic gene cluster, involves the assembly of an initial
polyketide macrolactam, followed by a series of post-PKS
modifications introducing a chlorine, two methyl groups,
a cyclic carbamate, an ester side chain, and an epoxide
function to form the ansamitocins. The unexpected isolation
of ansamitocinosides from A. pretiosum cultivated on solid
ISP2 medium implies the further potentials for post-PKS
modifications of maytansinoids and calls for alternative
strategies for natural product mining. The identification of
Asm25 as the N-Gtf adding glucose moiety to the macrolac-
tam amide nitrogen enriches the antibiotic glycotransferase,
especially the N-Gtf, toolbox. Notably, heterologous expres-
sion of asm25 in E. coli and solubilization of the Asm25
protein from the inclusion bodies provided for the first time
a recombinant antibiotic N-glycosyltransferase in soluble,
enzymatically active form. This allowed the in vitro
63–874, August 25, 2008 ª2008 Elsevier Ltd All rights reserved 871
Chemistry & Biology
Amide N-Glycosylation
characterization of Asm25 to examine the properties of this
enzyme and the catalytic mechanism of this unusual C-N
bond forming glycosylation. The broad substrate range of
the enzyme for different aglycones suggests the possibility
of using it to generate N-glucosylated derivatives of ansa-
mycins and other compounds for evaluation as improved
bioactive agents or potential prodrugs.
EXPERIMENTAL PROCEDURES
Materials, General Methods, and Instrumentation
Optical rotations were measured with a JASCO DIP-370 digital polarimeter in
MeOH solution. Mass spectra were measured on an API Qstar Pulsar spec-
trometer. NMR spectra were recorded on Bruker AM-400 and DRX-500
NMR spectrometers with TMS as internal standard. UDP-Glc, rifamycin, and
rifampicin were purchased from Sigma-Aldrich. Naphthomycin A was isolated
from the endophytic Streptomyces sp. CS (Lu and Shen, 2007). The prepara-
tion of PND-1-4, DDM, DDP-1, lebstatin, 17-DMLB, HC-14, and HC-52 is
described in the Supplemental Data. UDP-galactose, UDP-N-acetyl-glucos-
amine, UDP-glucuronic acid, ADP-Glc, and GDP-Glc were obtained as a gift
from Taifo Mahmud of the Department of Pharmaceutical Sciences, Oregon
State University, Corvallis. LC-MS for qualitative and quantitative analyses
was carried out using a Waters series HPLC 2695 instrument (Waters Corp.)
with a Thermo Finnigan LCQ Advantage (Thermo Finnigan) electrospray ioni-
zation mass detector (ion trap). For methods of LC-MS, see Supplemental
Data.
Strains, Plasmids, Culture Techniques, and Media
Actinosynnema pretiosum ssp. aurantium ATCC 31565 and its derivatives, and
Streptomyces hygroscopicus XM201 were grown either on solid or in liquid
ISP2 medium (containing 0.4% yeast extract, 1% malt extract, and 0.4%
glucose [pH 7.3]) or liquid TSBY medium [30 g/l tryptic soy broth (LabM; Topley
House), 10 g/l yeast extract, and 103 g/l sucrose]. Fermentations were carried
out for 7 days at 28�C. Subsequent extractions and purifications were
performed as described in the Supplemental Data. E. coli DH10B (Invitrogen),
ET12567(pUZ8002) (MacNeil et al., 1992), and BL21(DE3)pLysE (Invitrogen)
were used as hosts for plasmid construction, E. coli-Actinosynnema biparental
conjugation, and protein overexpression, respectively. The yeast Filobasidium
unigutulatum IFO0699 was used as indicator strain for ansamitocin and ansa-
mitocinoside bioassay, and cytotoxicity assays were performed on a panel of
HepG2 cells. pBluescript KS(�) and pET26a were used for plasmid construc-
tion. pHGF9053 (Minagawa et al., 2007) was the vector used for the gene in-
activation of asm25 and pOJ2600, which is a derivative of pOJ260 (Bierman
et al., 1992) with the original KpnI site removed, was used for asm10 inactiva-
tion. Integrative plasmid pIB139 (Wilkinson et al., 2002) was used for the com-
plementation of the asm25 mutant. pRSET-B (Invitrogen) was used as vector
for protein overexpression in E. coli. For the complete list of strains and
plasmids used in this article, see Table S5.
Isolation and Identification of Ansamitocinosides
The strain A. pretiosum ATCC 31565 was inoculated on a slope of solid ISP2
medium in a test tube and cultivated for 5 days at 28�C to allow seed culture.
Solid-state fermentation was performed with solid ISP2 medium (3 L, �150
petri dishes) for 7 days at 28�C. The culture was extracted five times with
EtOAc-MeOH-AcOH (80:15:5, v/v/v) to afford a crude extract of 21.0 g. By
following the isolation procedure described previously (Ma et al., 2007), three
ansamitocin glucosides, AGP-1 (5.0 mg), AGP-2 (8.0 mg), and AGP-3 (3.0 mg),
were obtained. The NMR and MS data of AGP-1 and AGP-2 were identical with
the literature (Ma et al., 2007). The ESI-MS and NMR (1H, 13C, DEPT, HSQC
and HMBC) spectra (Figures S2A–L), HPLC comparison profile (Figure S2M),
and NMR assignments (Tables S1A and S1B) for in vivo and in vitro AGP-3
are given in the Supplemental Data.
Inactivation of Glycosyltransferase Gene asm25
A 6646 bp SacI fragment containing asm25 was cleaved from cosmid 19F11
(Yu et al., 2002) of the ATCC 31565 genomic library and cloned into pBluescript
KS(�) digested with SacI, resulting in pJTU476. The second plasmid,
872 Chemistry & Biology 15, 863–874, August 25, 2008 ª2008 Elsev
pJTU477, was constructed through the ligation of the 2742 bp SacI-PvuII
and 3439 bp PvuII-SacI fragments of pJTU476 with SacI-digested pET26a.
Transfer of the 6181 bp SacI fragment from pJTU477 to SacI-digested
pHGF9053 generated pJTU478 for the subsequent inactivation of asm25.
pJTU478 was introduced into ATCC 31565 via E. coli-Actinosynnema biparen-
tal conjugation as described (Kieser et al., 2000). Exconjugants were selected
with 15 mg/ml thiostrepton. For the screening for double-crossover mutants,
an exconjugant was inoculated into liquid TSBY medium and cultivated for
three rounds without the presence of thiostrepton. Then the thiostrepton-sen-
sitive derivatives were analyzed through PCR amplification with a pair of
primers (Asm25-Det-F: 50-CCCCAGCACGGAGGAAGA-30 and Asm25-Det-
R: 50-AGCGGAGGAGGAGACCCA-30). The PCR amplification was done in
a thermocycler (Thermo, MBS Satellite 0.2) under the following conditions:
35 cycles of 30 s at 94�C, 30 s at 50.5�C, and 90 s at 72�C. The asm25 mutant,
named BLQ11, was also analyzed by fermentation and LC-MS analysis.
Complementation of Mutant BLQ11 with Cloned asm25
A 1.55 kb DNA fragment containing the asm25 gene was cleaved from
pJTU812 by NdeI and EcoRI and cloned into plasmid pIB139 to obtain
pJTU813, which was introduced from E. coli into BLQ11 through conjugation.
Confirmation was carried out through plasmid isolation from the exconjugants,
plasmid transformation of E. coli, and comparison between the newly purified
pJTU813 and the original one by restriction enzyme digestions. An exconju-
gant carrying the correct plasmid was named as BLQ15 and characterized
by fermentation and LC-MS analysis.
Cloning and Heterologous Expression of Recombinant
His6-tagged Asm25
Gene asm25 was amplified from the ATCC 31565 genome by PCR using
primers asm25-F (50-GGATCCACATATGCGCGTTCTGTTCACC-30, engi-
neered BamHI and NdeI sites underlined) and asm25-R (50-GAATTCACAC
CACCGCGACGCTC-30, engineered EcoRI site underlined). The amplified
gene fragment was digested with BamHI and EcoRI and ligated with
pRSET-B digested with the same enzymes to generate pJTU812. The se-
quence of the cloned asm25 was confirmed by DNA sequencing. pJTU812
was transformed into E. coli BL21(DE3)pLysE, and transformants were grown
in LB medium supplemented with ampicillin (100 mg/ml) and chloramphenicol
(17.5 mg/ml) at 37�C and 230 rpm for 12 h, then diluted 1:10 with fresh LB me-
dium. The diluted cultures were grown to OD600 = 0.4�0.6, then induced with
0.1�0.8 mM IPTG and incubated for 5 h at 37�C, 230 rpm.
Purification and Renaturation of Asm25 from Inclusion Bodies
E. coli cells (200 ml) induced for 5 h at 37�C were centrifuged at 8,000 g for
15 min and the pellet was resuspended in 3 ml renaturation buffer I (50 mM
Tris-HCl [pH 8.0], 300 mM NaCl, 10 mM MgCl2). Cells were then frozen in liquid
nitrogen, rapidly thawed in 37�C water for three cycles, and then sonicated on
ice six times for 15 s each at 15 s intervals. After centrifugation of the lysate at
4�C and 30,000 g for 20 min, the inclusion body pellet was resuspended in 3.0
ml renaturation buffer IIa (50 mM Tris-HCl [pH 8.0], 300 mM NaCl, 10 mM
MgCl2, 0.5% Triton, 2.0 M urea) for 5 min at room temperature and recovered
by centrifugation at 4�C and 15,000 g for 15 min. The same procedure was re-
peated twice in 3.0 ml renaturation buffer IIb (50 mM Tris-HCl [pH 8.0], 300 mM
NaCl, 10 mM MgCl2, 0.5% Triton). The purified Asm25 inclusion bodies were
suspended in 4.0 ml renaturation buffer III (50 mM Tris-HCl [pH 8.0], 300 mM
NaCl, 10 mM MgCl2, 8.0 M urea) and centrifuged at 8,000 g for 15 min; the pel-
let was then discarded. The supernatant was dialyzed against renaturation
buffer IV (50 mM Tris-HCl [pH 8.0], 10% glycerol, 100 mM NaCl, 0.1 M glycine,
0.5 mM EDTA, 1.0 mM DDT, 10 mM MgCl2) for 24 h at 4�C. After centrifugation
for 15 min at 4�C and 20,000 g, the supernatant (12.0 ml) was collected and
used for subsequent enzyme assays. Protein purity and molecular mass
were examined by SDS-PAGE followed by Coomassie Brilliant Blue staining
of the gel. For SDS-PAGE, the protein solution was mixed with Laemmli sam-
ple buffer containing b-mercaptoethanol (5%) and incubated at 40�C for
30 min. The protein concentration was measured to be 1.6 mM by the Bradford
method or by staining intensity compared with BSA concentration standards.
The yield of Asm25 was calculated to be�4.38 mg/l of induced cells. To make
a basal reaction solution for activity assay and storage at 4�C, glycerol and
ier Ltd All rights reserved
Chemistry & Biology
Amide N-Glycosylation
BSA were added to the enzyme solution in buffer IV at final concentrations of
30% and 1%, respectively.
SUPPLEMENTAL DATA
Supplemental Data include five figures, five tables, Supplemental Experimen-
tal Procedures, and Supplemental References, and can be found with this
article online at http://www.chembiol.com/cgi/content/full/15/8/863/DC1/.
ACKNOWLEDGMENTS
We thank Taifo Mahmud at Oregon State University, Corvallis, for providing
UDP-galactose, UDP-N-acetyl-glucosamine, UDP-glucuronic acid, ADP-Glc,
and GDP-Glc. This work was partially supported by the National Natural
Science Fund for Distinguished Young Scholars to Y. Shen (30325044), the
Natural Science Foundation of China (30600005, 30570019), and the Ministry
of Science and Technology (973 and 863 Programs).
Received: February 14, 2008
Revised: June 10, 2008
Accepted: June 13, 2008
Published: August 22, 2008
REFERENCES
Ahlert, J., Shepard, E., Lomovskaya, N., Zazopoulos, E., Staffa, A., Bachmann,
B.O., Huang, K., Fonstein, L., Czisny, A., Whitwam, R.E., et al. (2002). The cal-
icheamicin gene cluster and its iterative type I enediyne PKS. Science 297,
1173–1176.
Bierman, M., Logan, R., O’Brien, K., Seno, E.T., Rao, R.N., and Schoner, B.E.
(1992). Plasmid cloning vectors for the conjugal transfer of DNA from Escher-
ichia coli to Streptomyces spp. Gene 116, 43–49.
Brazier-Hicks, M., Offen, W.A., Gershater, M.C., Revett, T.J., Lim, E.K.,
Bowles, D.J., Davies, G.J., and Edwards, R. (2007). Characterization and en-
gineering of the bifunctional N- and O-glucosyltransferase involved in xenobi-
otic metabolism in plants. Proc. Natl. Acad. Sci. USA 104, 20238–20243.
Cassady, J.M., Chan, K.K., Floss, H.G., and Leistner, E. (2004). Recent devel-
opments in the maytansinoid antitumor agents. Chem. Pharm. Bull. (Tokyo) 52,
1–26.
Clark, E.D.B. (1998). Refolding of recombinant proteins. Curr. Opin. Biotech-
nol. 9, 157–163.
Fischbach, M.A., Lin, H., Liu, D.R., and Walsh, C.T. (2005). In vitro character-
ization of IroB, a pathogen-associated C-glycosyltransferase. Proc. Natl.
Acad. Sci. USA 102, 571–576.
Gao, Q., Zhang, C., Blanchard, S., and Thorson, J.S. (2006). Deciphering indo-
locarbazole and enediyne aminodideoxypentose biosynthesis through com-
parative genomics: insights from the AT2433 biosynthetic locus. Chem. Biol.
13, 733–743.
Gekko, K., and Timasheff, S.N. (1981). Mechanism of protein stabilization by
glycerol: preferential hydration in glycerol-water mixtures. Biochemistry 20,
4667–4676.
Ghisalba, O., and Nuesch, J. (1981). A genetic approach to the biosynthesis of
the rifamycin-chromophore in Nocardia mediterranei. IV. Identification of
3-amino-5-hydroxybenzoic acid as a direct precursor of the seven-carbon
amino starter-unit. J. Antibiot. (Tokyo) 34, 64–71.
Hatano, K., Akiyama, S., Asai, M., and Rickards, R.W. (1982). Biosynthetic or-
igin of aminobenzenoid nucleus (C7N-unit) of ansamitocin, a group of novel
maytansinoid antibiotics. J. Antibiot. (Tokyo) 35, 1415–1417.
Higashide, E., Asai, M., Ootsu, K., Tanida, S., Kozai, Y., Hasegawa, T., Kishi,
T., Sugino, Y., and Yoneda, M. (1977). Ansamitocin, a group of novel may-
tansinoid antibiotics with antitumour properties from Nocardia. Nature 270,
721–722.
Hoffmeister, D., Ichinose, K., Domann, S., Faust, B., Trefzer, A., Drager, G.,
Kirschning, A., Fischer, C., Kunzel, E., Bearden, D., et al. (2000). The NDP-
sugar co-substrate concentration and the enzyme expression level influence
Chemistry & Biology 15, 8
the substrate specificity of glycosyltransferases: cloning and characterization
of deoxysugar biosynthetic genes of the urdamycin biosynthetic gene cluster.
Chem. Biol. 7, 821–831.
Hyun, C.G., Bililign, T., Liao, J., and Thorson, J.S. (2003). The biosynthesis of
indolocarbazoles in a heterologous E. coli host. Chembiochem 4, 114–117.
Izawa, M., Tanida, S., and Asai, M. (1981). Ansamitocin analogs from a mutant
strain of Nocardia. II. Isolation and structure. J. Antibiot. (Tokyo) 34, 496–506.
Jackson, R.G., Lim, E.K., Li, Y., Kowalczyk, M., Sandberg, G., Hoggett, J.,
Ashford, D.A., and Bowles, D.J. (2001). Identification and biochemical charac-
terization of an Arabidopsis indole-3-acetic acid glucosyltransferase. J. Biol.
Chem. 276, 4350–4356.
Kato, Y., Bai, L., Xue, Q., Revill, W.P., Yu, T.W., and Floss, H.G. (2002). Func-
tional expression of genes involved in the biosynthesis of the novel polyketide
chain extension unit, methoxymalonyl-acyl carrier protein, and engineered
biosynthesis of 2-desmethyl-2-methoxy-6-deoxyerythronolide B. J. Am.
Chem. Soc. 124, 5268–5269.
Kibby, J.J., McDonald, I.A., and Rickards, R.W. (1980). 3-Amino-5-hydroxy-
benzoic acid as a key intermediate in ansamycin and maytansinoid biosynthe-
sis. J. Chem. Soc. Chem. Commun., 768–769.
Kieser, T., Bibb, M.J., Buttner, M.J., Chater, K.F., and Hopwood, D.A. (2000).
Practical Streptomyces Genetics (Norwich, U.K.: John Innes Foundation
Press).
Kim, C.G., Yu, T.W., Fryhle, C.B., Handa, S., and Floss, H.G. (1998). 3-Amino-
5-hydroxybenzoic acid synthase, the terminal enzyme in the formation of the
precursor of mC7N units in rifamycin and related antibiotics. J. Biol. Chem.
273, 6030–6040.
Kim, S.Y., Park, J.S., Chae, C.S., Hyun, C.G., Choi, B.W., Shin, J., and Oh, K.B.
(2007). Genetic organization of the biosynthetic gene cluster for the indolocar-
bazole K-252a in Nonomuraea longicatena JCM 11136. Appl. Microbiol.
Biotechnol. 75, 1119–1126.
Kupchan, S.M., Komoda, Y., Branfman, A.R., Sneden, A.T., Court, W.A.,
Thomas, G.J., Hintz, H.P., Smith, R.M., Karim, A., Howie, G.A., et al. (1977).
The maytansinoids. Isolation, structural elucidation, and chemical interrelation
of novel ansa macrolides. J. Org. Chem. 42, 2349–2357.
Lairson, L.L., Henrissat, B., Davies, G.J., and Withers, S.G. (2008). Glycosyl-
transferases: structures, functions, and mechanisms. Annu. Rev. Biochem.
77, 521–555.
Lu, C., and Shen, Y. (2007). A novel ansamycin, naphthomycin K from Strep-
tomyces sp. J. Antibiot. (Tokyo) 60, 649–653.
Lu, C., Bai, L., and Shen, Y. (2004). A novel amide N-glycoside of ansamitocins
from Actinosynnema pretiosum. J. Antibiot. (Tokyo) 57, 348–350.
Ma, J., Zhao, P.J., and Shen, Y.M. (2007). New amide N-glycosides of ansami-
tocins identified from Actinosynnema pretiosum. Arch. Pharm. Res. 30, 670–
673.
Mackenzie, P.I. (1990). Expression of chimeric cDNAs in cell culture defines
a region of UDP-glucuronosyltransferase involved in substrate selection.
J. Biol. Chem. 265, 3432–3435.
MacNeil, D.J., Gewain, K.M., Ruby, C.L., Dezeny, G., Gibbons, P.H., and
MacNeil, T. (1992). Analysis of Streptomyces avermitilis genes required for
avermectin biosynthesis utilizing a novel integration vector. Gene 111, 61–68.
Marangoni, A.G. (2003). Enzyme Kinetics: A Modern Approach (Hoboken, NJ:
John Wiley).
Minagawa, K., Zhang, Y., Ito, T., Bai, L., Deng, Z., and Mahmud, T. (2007).
ValC, a new type of C7-Cyclitol kinase involved in the biosynthesis of the an-
tifungal agent validamycin A. Chembiochem 8, 632–641.
Morisaki, N., Kobayashi, H., Iwasaki, S., Furihata, K., Dabbs, E.R., Yazawa, K.,
and Mikami, Y. (1995). Structure determination of ribosylated rifampicin and its
derivative: new inactivated metabolites of rifampicin by mycobacterial strains.
J. Antibiot. (Tokyo) 48, 1299–1303.
Mosmann, T. (1983). Rapid colorimetric assay for cellular growth and survival:
application to proliferation and cytotoxicity assays. J. Immunol. Methods 65,
55–63.
Moss, S.J., Bai, L., Toelzer, S., Carroll, B.J., Mahmud, T., Yu, T.W., and Floss,
H.G. (2002). Identification of asm19 as an acyltransferase attaching the
63–874, August 25, 2008 ª2008 Elsevier Ltd All rights reserved 873
Chemistry & Biology
Amide N-Glycosylation
biologically essential ester side chain of ansamitocins using N-desmethyl-4,
5-desepoxymaytansinol, not maytansinol, as its substrate. J. Am. Chem. Soc.
124, 6544–6545.
Mulichak, A.M., Losey, H.C., Walsh, C.T., and Garavito, R.M. (2001). Structure
of the UDP-glucosyltransferase GtfB that modifies the heptapeptide aglycone
in the biosynthesis of vancomycin group antibiotics. Structure 9, 547–557.
Nishizawa, T., Aldrich, C.C., and Sherman, D.H. (2005). Molecular analysis of
the rebeccamycin L-amino acid oxidase from Lechevalieria aerocolonigenes
ATCC 39243. J. Bacteriol. 187, 2084–2092.
Ohuchi, T., Ikeda-Araki, A., Watanabe-Sakamoto, A., Kojiri, K., Nagashima,
M., Okanishi, M., and Suda, H. (2000). Cloning and expression of a gene en-
coding N-glycosyltransferase (ngt) from Saccarothrix aerocolonigenes
ATCC39243. J. Antibiot. (Tokyo) 53, 393–403.
Onaka, H., Taniguchi, S., Igarashi, Y., and Furumai, T. (2002). Cloning of the
staurosporine biosynthetic gene cluster from Streptomyces sp. TP-A0274
and its heterologous expression in Streptomyces lividans. J. Antibiot. (Tokyo)
55, 1063–1071.
Onaka, H., Taniguchi, S., Igarashi, Y., and Furumai, T. (2003a). Characteriza-
tion of the biosynthetic gene cluster of rebeccamycin from Lechevalieria aero-
colonigenes ATCC 39243. Biosci. Biotechnol. Biochem. 67, 127–138.
Onaka, H., Taniguchi, S., Ikeda, H., Igarashi, Y., and Furumai, T. (2003b).
pTOYAMAcos, pTYM18, and pTYM19, actinomycete-Escherichia coli inte-
grating vectors for heterologous gene expression. J. Antibiot. (Tokyo) 56,
950–956.
Ootsu, K., Kozai, Y., Takeuchi, M., Ikeyama, S., Igarashi, K., Tsukamoto, K.,
Sugino, Y., Tashira, T., Tsukagoshi, S., and Sakurai, Y. (1980). Effects of
new antimitotic antibiotics, ansamitocins, on the growth of murine tumors
in vivo and on the assembly of microtubules in vitro. Cancer Res. 40, 1707–1717.
Price, N.C., and Stevens, L. (1982). Fundamental of Enzymology (Oxford:
Oxford University Press).
Rinehart, K.L., Jr., and Shield, L.S. (1976). Chemistry of the ansamycin antibi-
otics. Fortschr. Chem. Org. Naturst. 33, 231–307.
Salas, A.P., Zhu, L., Sanchez, C., Brana, A.F., Rohr, J., Mendez, C., and Salas,
J.A. (2005). Deciphering the late steps in the biosynthesis of the anti-tumour
indolocarbazole staurosporine: sugar donor substrate flexibility of the StaG
glycosyltransferase. Mol. Microbiol. 58, 17–27.
Sanchez, C., Butovich, I.A., Brana, A.F., Rohr, J., Mendez, C., and Salas, J.A.
(2002). The biosynthetic gene cluster for the antitumor rebeccamycin: char-
acterization and generation of indolocarbazole derivatives. Chem. Biol. 9,
519–531.
874 Chemistry & Biology 15, 863–874, August 25, 2008 ª2008 Elsev
Sanchez, C., Zhu, L., Brana, A.F., Salas, A.P., Rohr, J., Mendez, C., and Salas,
J.A. (2005). Combinatorial biosynthesis of antitumor indolocarbazole com-
pounds. Proc. Natl. Acad. Sci. USA 102, 461–466.
Sanchez, C., Mendez, C., and Salas, J.A. (2006). Indolocarbazole natural prod-
ucts: occurrence, biosynthesis, and biological activity. Nat. Prod. Rep. 23,
1007–1045.
Sensi, P., Margalith, P., and Timbal, M.T. (1959). Rifomycin, a new antibiotic:
preliminary report. Farmaco 14, 146–147.
Sneden, A.T., and Beemsterboer, G.L. (1980). Normaytansine, a new antileu-
kemic ansa macrolide from Maytenus buchananii. J. Nat. Prod. 43, 637–640.
Spiteller, P., Bai, L., Shang, G., Carroll, B.J., Yu, T.W., and Floss, H.G. (2003).
The post-polyketide synthase modification steps in the biosynthesis of the an-
titumor agent ansamitocin by Actinosynnema pretiosum. J. Am. Chem. Soc.
125, 14236–14237.
Stead, P., Latif, S., Blackaby, A.P., Sidebottom, P.J., Deakin, A., Taylor, N.L.,
Life, P., Spaull, J., Burrell, F., Jones, R., et al. (2000). Discovery of novel ansa-
mycins possessing potent inhibitory activity in a cell-based oncostatin M sig-
nalling assay. J. Antibiot. (Tokyo) 53, 657–663.
Walsh, C., Freel Meyers, C.L., and Losey, H.C. (2003). Antibiotic glycosyltrans-
ferases: antibiotic maturation and prospects for reprogramming. J. Med.
Chem. 46, 3425–3436.
Widdison, W.C., Wilhelm, S.D., Cavanagh, E.E., Whiteman, K.R., Leece, B.A.,
Kovtun, Y., Goldmacher, V.S., Xie, H., Steeves, R.M., Lutz, R.J., et al. (2006).
Semisynthetic maytansine analogues for the targeted treatment of cancer. J.
Med. Chem. 49, 4392–4408.
Wilkinson, C.J., Hughes-Thomas, Z.A., Martin, C.J., Bohm, I., Mironenko, T.,
Deacon, M., Wheatcroft, M., Wirtz, G., Staunton, J., and Leadlay, P.F.
(2002). Increasing the efficiency of heterologous promoters in actinomycetes.
J. Mol. Microbiol. Biotechnol. 4, 417–426.
Yang, X.A., Dong, X.Y., Li, Y., Wang, Y.D., and Chen, W.F. (2004). Purification
and refolding of a novel cancer/testis antigen BJ-HCC-2 expressed in the in-
clusion bodies of Escherichia coli. Protein Expr. Purif. 33, 332–338.
Yu, T.W., Bai, L., Clade, D., Hoffmann, D., Toelzer, S., Trinh, K.Q., Xu, J., Moss,
S.J., Leistner, E., and Floss, H.G. (2002). The biosynthetic gene cluster of the
maytansinoid antitumor agent ansamitocin from Actinosynnema pretiosum.
Proc. Natl. Acad. Sci. USA 99, 7968–7973.
Zhang, C., Albermann, C., Fu, X., Peters, N.R., Chisholm, J.D., Zhang, G., Gil-
bert, E.J., Wang, P.G., Van Vranken, D.L., and Thorson, J.S. (2006). RebG- and
RebM-catalyzed indolocarbazole diversification. Chembiochem 7, 795–804.
ier Ltd All rights reserved